Hydrogen-doped pipeline detection method based on eddy current and magnetic flux leakage composite detection

Abstract

The invention provides a hydrogen-doped pipeline detection method based on eddy current and magnetic flux leakage composite detection, and relates to the technical field of hydrogen-doped pipeline detection.The method comprises the steps that a reference magnetic field distribution vector and a reference electromagnetic induction response matrix of a hydrogen-doped pipeline are obtained, and a composite detection reference field model is constructed; executing leakage magnetic field response acquisition and abnormal field reconstruction under hydrogen-doped pipeline defect disturbance, extracting an abnormal magnetic field gradient change set, and generating a defect induced magnetic field disturbance distribution diagram; hydrogen-doped pipeline near-surface micro-crack induced eddy current response disturbance identification is executed, a near-surface current density abnormal change sequence is extracted, and a crack disturbance response feature map is generated; and performing multi-source disturbance feature fusion discrimination on the abnormal magnetic field gradient change set and the near-surface current density abnormal change sequence, and outputting a defect distribution result and a defect grade evaluation result of the hydrogen-doped pipeline, so that full-scale defect identification of the hydrogen-doped pipeline is realized through eddy current and magnetic flux leakage composite detection, working condition disturbance is effectively inhibited, and the detection accuracy is improved. And the defect detection accuracy and reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of hydrogen-doped pipeline inspection technology, and in particular to a method for inspecting hydrogen-doped pipelines based on a combined eddy current and magnetic flux leakage detection. Background Technology

[0002] Hydrogen-blended natural gas transportation is the core route for the large-scale storage, transportation, and consumption of hydrogen energy, and high-pressure pipelines are the key carriers for transporting hydrogen-blended media. However, during long-term service, hydrogen molecule penetration can cause hydrogen-induced damage to steel, resulting in a complex defect where near-surface microcrack initiation and propagation coexist with deep metal cross-sectional loss. At the same time, it accelerates the pipeline corrosion process, significantly increases the fatigue crack propagation rate and failure risk, and directly threatens the safe operation of the pipeline.

[0003] Among existing pipeline non-destructive testing technologies, magnetic flux leakage testing alone is effective in identifying deep wall thickness losses, but its sensitivity in detecting near-surface microcracks is insufficient. Eddy current testing alone can accurately capture near-surface defects, but it cannot effectively identify deep cross-sectional losses, making it difficult to achieve full-scale defect coverage detection in hydrogen-doped pipelines. Furthermore, existing composite testing schemes lack a unified reference field model and are susceptible to disturbances from operating conditions such as lift-off fluctuations, temperature drift, and differences in pipeline geometry and material conditions, leading to misjudgments and missed detections. These methods cannot meet the integrity assessment and safe operation and maintenance requirements for long-term service of hydrogen-doped pipelines.

[0004] Therefore, it is necessary to provide a hydrogen-doped pipeline detection method based on a combination of eddy current and magnetic flux leakage detection to solve the above-mentioned technical problems. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a hydrogen-doped pipeline inspection method based on eddy current and magnetic flux leakage composite detection. This method solves the problems that existing single detection methods cannot simultaneously identify near-surface microcracks and deep metal cross-sectional losses in hydrogen-doped pipelines across the entire scale, and that composite detection lacks a unified benchmark, is susceptible to operating condition disturbances, and suffers from insufficient accuracy and reliability in defect identification.

[0006] The present invention provides a method for detecting hydrogen-doped pipelines based on a combined eddy current and magnetic flux leakage detection method, the method comprising: Obtain the reference magnetic field distribution vector and reference electromagnetic induction response matrix of the hydrogen-doped pipeline under complete state, and construct a composite detection reference field model of the hydrogen-doped pipeline. Based on the composite detection reference field model, the leakage magnetic field response acquisition and abnormal field reconstruction under the defect disturbance of hydrogen-doped pipeline are performed, the abnormal magnetic field gradient change set is extracted, and the defect-induced magnetic field disturbance distribution map is generated. Based on the composite detection reference field model, the eddy current response disturbance identification induced by near-surface microcracks in hydrogen-doped pipelines is performed, the abnormal change sequence of near-surface current density is extracted, and the crack disturbance response feature map is generated. The set of abnormal magnetic field gradient changes and the sequence of abnormal near-surface current density changes are fused and judged by multi-source perturbation features, and the defect distribution results and defect level assessment results of hydrogen-doped pipelines are output.

[0007] Preferably, the step of obtaining the reference magnetic field distribution vector and reference electromagnetic induction response matrix of the hydrogen-doped pipeline in its complete state, and constructing a composite detection reference field model for the hydrogen-doped pipeline, includes the following steps: Obtain the base material information, geometric parameters and electromagnetic parameters of the material in the complete state of the hydrogen-doped pipeline. Establish equidistant detection paths along the axial direction of the hydrogen-doped pipeline and divide them into a set of axial discrete detection units. Calculate the equivalent permeability of each axial discrete detection unit. Based on the set of axial discrete detection units and the equivalent permeability of each axial discrete detection unit, combined with the magnetization device parameters and probe lift-off height, the reference magnetic induction intensity of each axial discrete detection unit under the complete state of the hydrogen-doped pipeline is calculated, and the reference magnetic field distribution vector is generated. Based on the set of axial discrete detection units and the equivalent permeability of each axial discrete detection unit, combined with the high-frequency excitation parameters and detection coil parameters, the standard skin depth and reference induced voltage amplitude of each axial discrete detection unit under the complete state of the hydrogen-doped pipeline are calculated, and the reference electromagnetic induction response matrix is ​​generated. Based on the reference magnetic field distribution vector and the reference electromagnetic induction response matrix, an excitation scan of the hydrogen-doped pipeline in its complete state is performed, the composite reference response value of each axial discrete detection unit is calculated, and a composite detection reference field model containing the reference magnetic field distribution vector and the reference electromagnetic induction response matrix is ​​generated.

[0008] Preferably, the step of acquiring the leakage magnetic field response and reconstructing the abnormal field under the disturbance of hydrogen-doped pipeline defects based on the composite detection reference field model, extracting the abnormal magnetic field gradient change set, and generating a defect-induced magnetic field disturbance distribution map specifically includes: Based on the reference magnetic field distribution vector and the set of axial discrete detection units, the working parameters of the magnetization device and the magnetic sensing unit are configured to apply a constant magnetization field to the hydrogen-doped pipeline until the wall of the hydrogen-doped pipeline enters a stable near-saturation magnetization range. Along the detection path of the hydrogen-doped pipeline, the magnetic sensing unit collects the three components of leakage magnetic field signals in the axial, radial and tangential directions of each axial discrete detection unit. Combined with the detection propulsion speed and sampling time, the axial sampling position corresponding to each sampling point is calculated. Based on the three components of leakage magnetic field signals in the axial, radial and tangential directions of each axial discrete detection unit, the measured total leakage magnetic field amplitude of each axial discrete detection unit is synthesized. The measured total leakage magnetic flux amplitude of each axial discrete detection unit is bound to the corresponding axial sampling position to form a measured leakage magnetic flux response sequence arranged continuously along the detection path.

[0009] Preferably, based on the measured leakage magnetic response sequence, the reference magnetic induction intensity corresponding to each axial discrete detection unit is retrieved, the measured-reference correspondence under the same position and working condition constraints is established, and the normalized abnormal offset value of each axial discrete detection unit is calculated by combining the detection working condition disturbance parameters. The formula for calculating the normalized outlier offset is as follows: ; In the formula, is the normalized anomaly offset value of the i-th axial discrete detection unit; The measured total leakage flux amplitude of the i-th axial discrete detection unit; The reference magnetic flux density is the i-th axial discrete detection unit. For the local calibration correction coefficient of the i-th axial discrete detection unit; The actual lift-off height between the magnetization module and the outer surface of the hydrogen-doped pipe at the i-th axial discrete detection unit; Let be the detection temperature of the i-th axial discrete detection unit; To calibrate the detection temperature; Let be the magnetization current of the i-th axial discrete detection unit; To calibrate the magnetizing current; This is the compensation coefficient; A local sliding window is set based on the normalized abnormal offset value, and the local offset ratio of each axial discrete detection unit is calculated. The formula for calculating the local offset ratio is as follows: ; In the formula, The local offset ratio of the i-th axial discrete detection unit; To prevent stable terms with a denominator of zero; This represents the number of axial discrete detection units within the local sliding window. It is the sum of the absolute values ​​of the normalized anomaly offset values ​​of each axial discrete detection unit within a local sliding window centered on the i-th axial discrete detection unit; The normalized abnormal offset value of each axial discrete detection unit is bound to the local offset ratio to generate an initial abnormal magnetic field offset sequence.

[0010] Preferably, based on the initial abnormal magnetic field offset sequence, the normalized abnormal offset values ​​of each axial discrete detection unit are sorted according to the axial position of the detection path. The first-order spatial gradient of the normalized anomaly offset value of each axial discrete detection unit is calculated using the central difference method. Based on the first-order spatial gradient, the second-order spatial gradient of the normalized anomaly offset value of each axial discrete detection unit is calculated. By combining the normalized anomaly offset value and the corresponding first-order spatial gradient, second-order spatial gradient and detection step size, the local gradient energy of each axial discrete detection unit is calculated to generate a candidate anomaly field gradient set.

[0011] Preferably, a set of candidate anomaly field gradients is obtained, and anomaly segments are selected based on local gradient energy. Smoothing fitting and peak search are performed on each abnormal section to determine the location parameters of the abnormal section. Combined with the local gradient energy of each axial discrete detection unit in the abnormal section, the comprehensive abnormal intensity of the abnormal section is calculated and the defect-induced magnetic field disturbance zone is screened out. The formula for calculating the overall anomaly intensity is as follows: ; In the formula, The overall anomaly intensity of the k-th anomalous segment; The number of axial discrete detection units contained in the k-th abnormal segment; Let be the local gradient energy of the i-th axial discrete detection unit; The axial position of the i-th axial discrete detection unit within the abnormal section; This is the center location of the k-th abnormal segment; The equivalent width of the k-th abnormal segment; All defect-induced magnetic field disturbance regions are mapped to the surface coordinate system of the hydrogen-doped pipeline to generate a set of abnormal magnetic field gradient changes and a distribution map of defect-induced magnetic field disturbances.

[0012] Preferably, the step of identifying eddy current response disturbances induced by near-surface microcracks in hydrogen-doped pipelines based on a composite detection reference field model, extracting anomalous changes in near-surface current density, and generating a crack disturbance response feature map specifically includes: Based on the reference electromagnetic induction response matrix and the set of axial discrete detection units, high-frequency electromagnetic excitation parameters and detection coil parameters are configured, and high-frequency alternating electromagnetic excitation is applied to the hydrogen-doped pipe along the detection path to form a skin layer induced current near the surface of the hydrogen-doped pipe. The measured equivalent resistance and measured equivalent inductance of each axial discrete detection unit are collected synchronously. The standard skin depth and measured equivalent impedance amplitude of each axial discrete detection unit are calculated. A measured eddy current complex response sequence is established that corresponds one-to-one with the measured equivalent impedance amplitude and the reference electromagnetic induction response matrix. Based on the standard skin depth and the measured equivalent impedance amplitude, the actual induced current density of each axial discrete detection unit is calculated by inversion, and a comparison sequence of the actual induced current density and the reference current density is generated. The formula for calculating the actual induced current density is as follows: ; In the formula, The actual induced current density of the i-th axial discrete detection unit; Let be the measured equivalent impedance amplitude of the i-th axial discrete detection unit; The measured induced voltage amplitude of the i-th axial discrete detection unit; The equivalent radius of the detection coil; Let be the standard skin depth of the i-th axial discrete detection unit; denoted as the measured phase lag of the i-th axial discrete detection unit.

[0013] Preferably, based on the comparison sequence of the actual induced current density and the reference current density, for each axial discrete detection unit, combined with the detection condition disturbance parameters, the normalized current density offset value and the corresponding first-order spatial gradient and second-order spatial gradient are calculated. By combining the normalized current density offset value and the corresponding first-order and second-order spatial gradients, the local anomaly energy of each axial discrete detection unit is calculated and candidate crack zones are obtained. The comprehensive crack intensity of each candidate crack zone is obtained and the near-surface crack disturbance zone is screened. The near-surface crack disturbance zone is mapped to the surface coordinate system of the hydrogen-doped pipeline to generate the near-surface current density anomaly change sequence and crack disturbance response feature map.

[0014] Preferably, the step of fusing and judging the multi-source perturbation features of the abnormal magnetic field gradient change set and the near-surface current density abnormal change sequence to output the defect distribution result and defect level assessment result of the hydrogen-doped pipeline specifically includes: Obtain the set of abnormal magnetic field gradient changes and the sequence of abnormal near-surface current density changes, and establish a unified coordinate system with the detection start point as the axial zero point and the cumulative displacement of the encoder as the axial reference; Map the magnetic flux leakage anomaly region and the eddy current anomaly region to a unified coordinate system, and calculate the axial overlap ratio of each group of magnetic flux leakage anomaly region and eddy current anomaly region. Calculate the center offset distance between each group of magnetic flux leakage anomaly zone and eddy current anomaly zone, and combine the axial overlap ratio to obtain the composite defect alignment set; For each group of composite defect regions in the composite defect alignment set, the perturbation consistency coefficient and composite confidence value are calculated by combining the axial overlap ratio and the center offset distance, and a composite defect confidence sequence is generated. The formula for calculating the disturbance consistency coefficient is as follows: ; In the formula, is the disturbance consistency coefficient between the 0th leakage magnetic anomaly region and the qth eddy current anomaly region; denoted as the axial overlap ratio between the 0th leakage magnetic anomaly region and the qth eddy current anomaly region; The center offset distance between the 0th magnetic flux leakage anomaly region and the qth eddy current anomaly region; The center offset attenuation length; is the normalized flux leakage intensity of the o-th flux leakage anomaly region; is the normalized intensity of the eddy current in the q-th eddy current anomaly region; The formula for calculating the composite credibility value is: ; In the formula, The composite confidence value is the result of the o-th magnetic flux leakage anomaly region and the q-th eddy current anomaly region forming a composite defect. This is the slope control coefficient; This is the confidence threshold.

[0015] Preferably, based on the composite defect confidence sequence, for each group of composite defect regions, the corresponding leakage magnetic flux normalized intensity, eddy current normalized intensity, composite confidence value and axial influence length are extracted, and the grade score value of each group of composite defect regions is calculated by weighting. Based on the rating score, the composite defects are classified and the risk level is determined to generate the defect level assessment results for the hydrogen-doped pipeline. The defect level assessment results are mapped to the surface coordinate system of the hydrogen-doped pipeline to generate the defect distribution results of the hydrogen-doped pipeline.

[0016] Compared with related technologies, the hydrogen-doped pipeline detection method based on eddy current and magnetic flux leakage combined detection provided by this invention has the following advantages: This invention constructs a composite detection reference field model for hydrogen-doped pipelines by acquiring the reference magnetic field distribution vector and reference electromagnetic induction response matrix under the intact state of the pipeline. Based on the composite detection reference field model, it performs leakage magnetic field response acquisition and abnormal field reconstruction under defect disturbances in the hydrogen-doped pipeline, extracts the abnormal magnetic field gradient change set, and generates a defect-induced magnetic field disturbance distribution map. Based on the composite detection reference field model, it performs eddy current response disturbance identification induced by near-surface microcracks in the hydrogen-doped pipeline, extracts the near-surface current density abnormal change sequence, and generates a crack disturbance response feature map. It performs multi-source disturbance feature fusion and discrimination by combining the abnormal magnetic field gradient change set and the near-surface current density abnormal change sequence, and outputs the defect distribution results and defect level assessment results of the hydrogen-doped pipeline. Thus, through the composite detection of eddy current and leakage magnetic flux, it achieves full-scale defect identification of hydrogen-doped pipelines, effectively suppresses operating condition disturbances, and significantly improves the accuracy and reliability of defect detection.

[0017] This invention constructs a composite eddy current and magnetic flux leakage detection reference field model, unifying the reference benchmarks for the two types of detection channels. This effectively suppresses the influence of operational disturbances such as lift-off fluctuations, temperature drift, and pipeline geometric differences, improving the benchmark accuracy for anomaly identification. The method of this invention achieves accurate identification of deep metal cross-sectional losses in pipelines through the magnetic flux leakage detection channel and achieves highly sensitive detection of near-surface hydrogen-induced microcracks through the eddy current detection channel. This invention achieves collaborative identification of deep cross-sectional losses and near-surface cracks through spatial coordinate fusion and consistency discrimination of multi-source disturbance features, significantly reducing the false positive and false negative rates of defects. Simultaneously, it establishes a quantitative assessment system for defect levels, which can directly output defect distribution results and risk levels, enabling integrity assessment and safe operation and maintenance of hydrogen-doped pipelines, significantly improving the safety and stability of hydrogen-doped pipeline operation. Attached Figure Description

[0018] Figure 1 The flowchart illustrates the hydrogen-doped pipeline detection method based on a combination of eddy current and magnetic flux leakage detection, as provided in this embodiment of the invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] like Figure 1 The diagram shown is a flowchart of a hydrogen-doped pipeline detection method based on a combined eddy current and magnetic flux leakage detection method provided in an embodiment of the present invention. Figure 1 The execution entity of the method shown can be a software and / or hardware device. The execution entity of this application can include, but is not limited to, at least one of the following: user equipment, network equipment, etc. User equipment can include, but is not limited to, computers, smartphones, personal digital assistants (PDAs), and the aforementioned electronic devices. Network equipment can include, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of computers or network servers. Cloud computing is a type of distributed computing, consisting of a super virtual computer composed of a group of loosely coupled computers. This embodiment does not limit this. Steps S1 to S4 are detailed as follows: S1. Obtain the reference magnetic field distribution vector and reference electromagnetic induction response matrix of the hydrogen-doped pipeline under complete state, and construct the composite detection reference field model of the hydrogen-doped pipeline. S2, based on the composite detection reference field model, performs leakage magnetic field response acquisition and abnormal field reconstruction under hydrogen-doped pipeline defect disturbance, extracts abnormal magnetic field gradient change set, and generates defect-induced magnetic field disturbance distribution map; S3, based on the composite detection reference field model, performs eddy current response disturbance identification induced by near-surface microcracks in hydrogen-doped pipelines, extracts the abnormal change sequence of near-surface current density, and generates crack disturbance response feature map; S4 performs multi-source perturbation feature fusion and discrimination by combining the set of abnormal magnetic field gradient changes with the sequence of abnormal near-surface current density changes, and outputs the defect distribution results and defect level assessment results of the hydrogen-doped pipeline.

[0021] Existing pipeline composite inspection technologies are mainly divided into three categories: system fusion, data fusion, and mechanism fusion. Among them, system fusion technologies often adopt a multi-inspection section separate deployment scheme, resulting in low system coordination. Data fusion technologies only focus on the fusion processing of back-end data, neglecting the physical mechanism coordination and mutual interference suppression between different inspection technologies, making it difficult to meet the high sensitivity detection requirements of hydrogen-doped pipelines for smaller-sized defects.

[0022] To address the aforementioned technical problems, this invention employs an integrated structure in hardware, where eddy current and magnetic flux leakage detection units are coaxially arranged along the detection path. Simultaneously, at the software level, advanced signal processing methods such as empirical mode decomposition are introduced to effectively separate and accurately extract magnetic flux leakage magnetization interference and eddy current detection signals. By leveraging the strong identification advantage of magnetic flux leakage detection for deep pipe wall thickness loss, combined with the high sensitivity of eddy current detection for near-surface microcracks, this invention achieves a deep mechanistic-level fusion of the two detection technologies. This significantly improves the detection capability of minute defects across all scales in hydrogen-doped pipes, effectively avoiding the shortcomings of existing composite detection schemes, such as poor synergy, susceptibility to interference, and easy omission of minute defects.

[0023] In practical applications, taking a complete and defect-free hydrogen-doped pipeline as a reference, an equidistant detection path is established along the pipeline axis and axial discrete detection units are divided. Combining pipeline base material information, geometric parameters, and material electromagnetic properties, the equivalent permeability and initial conductivity of each detection unit are calculated. The reference magnetic induction intensity of each unit under the complete pipeline state is solved to form a reference magnetic field distribution vector. Simultaneously, electromagnetic parameters such as the standard skin depth and reference induced voltage under the complete state are calculated to construct a reference electromagnetic induction response matrix. Finally, a composite detection reference field model is generated through normalization processing to establish a unified zero reference reference for the two types of detection channels: leakage magnetic field and eddy current. This can effectively suppress the interference of operating condition disturbances such as lift-off fluctuations, temperature drift, and differences in pipeline geometry and material state on the detection results.

[0024] Based on the composite detection reference field model, a constant magnetization field is applied to the pipeline to bring the pipe wall into a stable near-saturation magnetization range. The axial, radial, and tangential leakage magnetic field signals of the pipeline are collected along the detection path. The measured total leakage magnetic field amplitude is synthesized and a measured leakage magnetic field response sequence is formed. Through reference difference and normalization correction of operating parameters under the same location and operating conditions, the initial abnormal magnetic field offset is extracted, the spatial gradient and local gradient energy of the abnormal field are calculated, continuous abnormal sections are screened and spatial reconstruction is completed. Finally, the abnormal magnetic field gradient change set and the defect-induced magnetic field disturbance distribution map are generated.

[0025] Based on the composite detection reference field model, high-frequency alternating electromagnetic excitation is applied to the pipeline to form a skin layer induced current near the pipe wall surface. The impedance, induced voltage, and phase information of the detection coil are collected simultaneously to form a measured eddy current complex response sequence that corresponds one-to-one with the reference matrix. The actual induced current density distribution near the pipe surface is obtained through inversion calculation. Differential analysis and operating condition compensation are performed with the reference current density to extract the abnormal change sequence of the near surface current density. The local abnormal energy is calculated and candidate crack zones are screened. Finally, the abnormal change sequence of the near surface current density and the crack disturbance response feature map are generated.

[0026] Finally, the abnormal magnetic field gradient change set of the leakage magnetic field channel and the abnormal change sequence of near-surface current density of the eddy current channel are mapped to a unified spatial coordinate system. Composite candidate defect areas are screened through spatial overlap analysis and center offset verification. The disturbance consistency coefficient and composite credibility are calculated to complete the credibility ranking of composite defects. A defect level scoring system is constructed through multi-dimensional weighted fusion to realize defect type classification and risk level assessment. Finally, the defect distribution results and defect level assessment results of the entire hydrogen-doped pipeline are output, completing the defect integrity identification required for the full life cycle safety assessment of the hydrogen-doped pipeline.

[0027] The process of obtaining the reference magnetic field distribution vector and reference electromagnetic induction response matrix of the hydrogen-doped pipeline under its complete state, and constructing a composite detection reference field model for the hydrogen-doped pipeline, includes the following steps: Obtain the base material information, geometric parameters and electromagnetic parameters of the material in the complete state of the hydrogen-doped pipeline. Establish equidistant detection paths along the axial direction of the hydrogen-doped pipeline and divide them into a set of axial discrete detection units. Calculate the equivalent permeability of each axial discrete detection unit. Based on the set of axial discrete detection units and the equivalent permeability of each axial discrete detection unit, combined with the magnetization device parameters and probe lift-off height, the reference magnetic induction intensity of each axial discrete detection unit under the complete state of the hydrogen-doped pipeline is calculated, and the reference magnetic field distribution vector is generated. Based on the set of axial discrete detection units and the equivalent permeability of each axial discrete detection unit, combined with the high-frequency excitation parameters and detection coil parameters, the standard skin depth and reference induced voltage amplitude of each axial discrete detection unit under the complete state of the hydrogen-doped pipeline are calculated, and the reference electromagnetic induction response matrix is ​​generated. Based on the reference magnetic field distribution vector and the reference electromagnetic induction response matrix, an excitation scan of the hydrogen-doped pipeline in its complete state is performed, the composite reference response value of each axial discrete detection unit is calculated, and a composite detection reference field model containing the reference magnetic field distribution vector and the reference electromagnetic induction response matrix is ​​generated.

[0028] First, basic information such as the base material grade, heat treatment state, and geometric dimensions of the hydrogen-doped pipeline in its intact state, as well as core electromagnetic parameters such as initial permeability and initial conductivity, are collected. An equidistant detection path is established along the pipeline axis, and axial discrete detection units are formed according to the detection step length. Separate segments are performed for sections with different material microstructures, such as welds, heat-affected zones, and cold-bending zones, to avoid sharing reference parameters among different microstructure areas. Initial parameter calibration is completed through standard sample calibration and on-site no-load test scanning. Then, segmented weighted smoothing is performed to reduce parameter fluctuations caused by surface roughness, temperature drift, and probe lift-off perturbations. Finally, the equivalent permeability of each axial discrete detection unit is calculated.

[0029] Based on the set of axial discrete detection units and their corresponding equivalent permeability, combined with the core parameters such as the number of coil turns, excitation current, and yoke spacing of the magnetization device, as well as the actual lift-off height between the probe and the pipe surface, magnetic circuit attenuation correction is introduced on the basis of ideal magnetic circuit calculation. Taking into account factors such as wall thickness coupling, lift-off attenuation, and pipe diameter normalization, the reference magnetic induction intensity corresponding to each axial discrete detection unit under the condition of a complete and defect-free pipe is calculated, and the reference magnetic field distribution vector is generated after arranging them in the order of the detection path.

[0030] Based on the set of axial discrete detection units and their corresponding equivalent permeability, combined with parameters such as high-frequency excitation frequency, number of turns of the detection coil, effective area, and equivalent impedance, the standard skin depth of each detection unit is first calculated to clarify the effective range of the alternating electromagnetic field on the surface of the pipe. Then, considering factors such as magnetic field strength, lift-off attenuation, and coil impedance characteristics, the reference induced voltage amplitude of each detection unit in the complete state is calculated. Simultaneously, supporting parameters such as the standard induced current attenuation factor and the standard phase hysteresis are obtained, and a reference electromagnetic induction response matrix is ​​constructed according to the dimensions of detection unit and excitation frequency.

[0031] Based on the reference magnetic field distribution vector and the reference electromagnetic induction response matrix, a simulated scan or physical test scan of the pipeline under complete conditions is performed to form a complete reference magnetic flux sequence, reference induction amplitude and phase sequence. Through normalization processing, the two reference channels with different dimensions, leakage magnetic field and eddy current, are integrated into the same evaluation framework. The composite reference response value of each detection unit is calculated, and finally a composite detection reference field model containing the reference magnetic field distribution vector and the reference electromagnetic induction response matrix is ​​generated.

[0032] The process, based on a composite detection reference field model, involves acquiring the leakage magnetic field response and reconstructing the abnormal field under hydrogen-doped pipeline defect disturbances, extracting the abnormal magnetic field gradient change set, and generating a defect-induced magnetic field disturbance distribution map. Specifically, this includes: Based on the reference magnetic field distribution vector and the set of axial discrete detection units, the working parameters of the magnetization device and the magnetic sensing unit are configured to apply a constant magnetization field to the hydrogen-doped pipeline until the wall of the hydrogen-doped pipeline enters a stable near-saturation magnetization range. Along the detection path of the hydrogen-doped pipeline, the magnetic sensing unit collects the three components of leakage magnetic field signals in the axial, radial and tangential directions of each axial discrete detection unit. Combined with the detection propulsion speed and sampling time, the axial sampling position corresponding to each sampling point is calculated. Based on the three components of leakage magnetic field signals in the axial, radial and tangential directions of each axial discrete detection unit, the measured total leakage magnetic field amplitude of each axial discrete detection unit is synthesized. The measured total leakage magnetic flux amplitude of each axial discrete detection unit is bound to the corresponding axial sampling position to form a measured leakage magnetic flux response sequence arranged continuously along the detection path.

[0033] Based on the measured leakage magnetic response sequence, the reference magnetic induction intensity corresponding to each axial discrete detection unit is retrieved, and the measured-reference correspondence under the same position and working condition constraints is established. Combined with the detection working condition disturbance parameters, the normalized abnormal offset value of each axial discrete detection unit is calculated. The formula for calculating the normalized outlier offset is as follows: ; In the formula, is the normalized anomaly offset value of the i-th axial discrete detection unit; The measured total leakage flux amplitude of the i-th axial discrete detection unit; The reference magnetic flux density is the i-th axial discrete detection unit. For the local calibration correction coefficient of the i-th axial discrete detection unit; The actual lift-off height between the magnetization module and the outer surface of the hydrogen-doped pipe at the i-th axial discrete detection unit; Let be the detection temperature of the i-th axial discrete detection unit; To calibrate the detection temperature; Let be the magnetization current of the i-th axial discrete detection unit; To calibrate the magnetizing current; This is the compensation coefficient; A local sliding window is set based on the normalized abnormal offset value, and the local offset ratio of each axial discrete detection unit is calculated. The formula for calculating the local offset ratio is as follows: ; In the formula, The local offset ratio of the i-th axial discrete detection unit; To prevent stable terms with a denominator of zero; This represents the number of axial discrete detection units within the local sliding window. It is the sum of the absolute values ​​of the normalized anomaly offset values ​​of each axial discrete detection unit within a local sliding window centered on the i-th axial discrete detection unit; The normalized abnormal offset value of each axial discrete detection unit is bound to the local offset ratio to generate an initial abnormal magnetic field offset sequence.

[0034] Based on the reference magnetic field distribution vector in the composite detection reference field model and the set of axial discrete detection units, the working parameters of the magnetization device and the magnetic sensing unit are configured. The magnetization device is driven by a constant current to apply a continuous and stable magnetization field to the wall of the hydrogen-doped pipeline. Key parameters such as magnetization current and magnetic pole spacing are strictly controlled to bring the pipeline wall into a stable near-saturation magnetization range. This reduces the interference of differences in pipeline material structure and local remanent magnetization on the detection results, ensuring that the abnormal magnetic field response acquired later is mainly caused by magnetic circuit disturbances caused by defects. At the same time, the magnetization conditions are strictly matched to those during the modeling of the reference magnetic field.

[0035] Along a pre-defined equidistant detection path along the pipeline axis, magnetic sensing units composed of Hall arrays or magnetoresistive arrays simultaneously acquire the magnetic induction intensity signals of the axial, radial, and tangential orthogonal components of the leakage magnetic field of pipeline wall defects at each discrete axial detection unit location. Simultaneously, the instantaneous advance speed of the detection head is acquired in real time via a mileage wheel or encoder. The time-domain sampling data is rigorously mapped to the pipeline axial spatial coordinate system, calculating the accurate axial sampling position corresponding to each sampling point. This avoids distortion of the leakage magnetic field location caused by fluctuations in the detection advance speed, ensuring a one-to-one spatial binding between the measured signal and the discrete axial detection unit.

[0036] For the three-component leakage magnetic field signal collected by each axial discrete detection unit, the measured total leakage magnetic field amplitude corresponding to the detection unit is calculated by vector synthesis. In this way, the magnetic field information of the three orthogonal directions can be integrated into a single amplitude feature quantity, which not only completely preserves the overall intensity information of the defect leakage magnetic field, but also achieves the unification of the dimensions and representation form with the reference magnetic induction intensity in the reference magnetic field distribution vector.

[0037] The measured total leakage magnetic flux amplitude of each axial discrete detection unit is bound to its corresponding accurate axial sampling position, sampling time, and three-component raw data, and arranged continuously according to the axial order of the detection path, ultimately forming a measured leakage magnetic flux response sequence that is completely distributed along the pipeline axis. This sequence achieves a one-to-one correspondence with the index of the previously established reference magnetic field distribution vector, and also completely preserves the spatial distribution characteristics and temporal sampling information of the leakage magnetic flux signal.

[0038] Based on the generated measured magnetic flux leakage response sequence, the reference magnetic induction intensity of each detection unit in the composite detection reference field model is retrieved through the spatial index of the axial discrete detection unit. With the same spatial position, the same lift-off condition, and the same magnetization parameter as the core constraints, a one-to-one correspondence between the measured signal and the reference data is established to ensure the consistency of the differential comparison under the working conditions and to eliminate the systematic deviation caused by inherent factors such as pipeline geometry and pipe diameter changes.

[0039] Based on the measured-benchmark correspondence, and considering operational disturbance parameters such as lift-off height fluctuations, ambient temperature changes, and minute fluctuations in magnetization current during the detection process, the difference between the measured total leakage magnetic flux amplitude and the benchmark magnetic induction intensity is normalized and corrected. Specifically, a compensation coefficient is used to suppress the global signal deviation caused by various operational disturbances, eliminating signal changes caused by non-defect factors such as weld reinforcement height, probe bounce, and magnetization fluctuations. The abnormal magnetic field offset induced only by defects is extracted, resulting in the normalized abnormal offset value corresponding to each axial discrete detection unit. This solves the problem that direct differential cannot distinguish between operational disturbances and actual defect anomalies.

[0040] Based on the normalized anomaly offset values ​​of each detection unit, a local sliding window adapted to the detection step size is set. With each detection unit as the center of the window, the local offset ratio of that unit is calculated. The local offset ratio can evaluate the relative significance of the anomaly offset value of a single point in the neighborhood background, distinguish the sharp anomalies caused by local defects from the slow signal fluctuations caused by the overall material inhomogeneity of the pipeline, further enhance the abnormal characteristics of real defects, and suppress the interference of background noise and global disturbances.

[0041] The normalized anomaly offset value of each axial discrete detection unit is bound to its corresponding local offset ratio and spatial location index, and arranged according to the axial order of the detection path to generate an initial anomaly magnetic field offset sequence. This sequence fully quantifies the degree and local saliency of leakage magnetic field anomalies at various locations in the pipeline, preserving the spatial distribution characteristics of defect anomalies while also correcting for operating disturbances and suppressing background noise.

[0042] It should be noted that the local calibration correction coefficient ranges from 0.9 to 1.1, determined by fitting the results of on-site test scans of complete, defect-free pipe sections. A default value of 1.0 is used for straight pipe sections, while fine adjustments are made based on test scan results for sections with material microstructure differences, such as welds and heat-affected zones, to compensate for the baseline system deviation caused by the local electromagnetic characteristics of the pipe. All three types of compensation coefficients are designed to counteract the interference of operating condition disturbances on the leakage magnetic field signal and to isolate actual defect anomalies. The value of the zero-stability term should be slightly higher than the background noise peak value of the magnetic sensing unit to avoid calculation errors with a denominator of zero and to avoid interfering with the statistical evaluation of actual anomaly signals.

[0043] Based on the initial abnormal magnetic field offset sequence, the normalized abnormal offset values ​​of each axial discrete detection unit are sorted according to the axial position of the detection path. The first-order spatial gradient of the normalized anomaly offset value of each axial discrete detection unit is calculated using the central difference method. Based on the first-order spatial gradient, the second-order spatial gradient of the normalized anomaly offset value of each axial discrete detection unit is calculated. By combining the normalized anomaly offset value and the corresponding first-order spatial gradient, second-order spatial gradient and detection step size, the local gradient energy of each axial discrete detection unit is calculated to generate a candidate anomaly field gradient set.

[0044] Obtain the set of candidate anomaly field gradients, and filter out anomaly segments based on local gradient energy; Smoothing fitting and peak search are performed on each abnormal section to determine the location parameters of the abnormal section. Combined with the local gradient energy of each axial discrete detection unit in the abnormal section, the comprehensive abnormal intensity of the abnormal section is calculated and the defect-induced magnetic field disturbance zone is screened out. The formula for calculating the overall anomaly intensity is as follows: ; In the formula, The overall anomaly intensity of the k-th anomalous segment; The number of axial discrete detection units contained in the k-th abnormal segment; Let be the local gradient energy of the i-th axial discrete detection unit; The axial position of the i-th axial discrete detection unit within the abnormal section; This is the center location of the k-th abnormal segment; The equivalent width of the k-th abnormal segment; All defect-induced magnetic field disturbance regions are mapped to the surface coordinate system of the hydrogen-doped pipeline to generate a set of abnormal magnetic field gradient changes and a distribution map of defect-induced magnetic field disturbances.

[0045] Based on the initial abnormal magnetic field offset sequence, the normalized abnormal offset values ​​of each axial discrete detection unit are spatially sorted according to the axial position of the detection path to eliminate the misalignment between the sampling time sequence and the spatial position of the pipeline, and to establish a continuous spatial domain sequence of abnormal signals.

[0046] The central difference method is used to calculate the first-order spatial gradient of the normalized anomaly offset value of each axial discrete detection unit. This characterizes the rate of change of the leakage magnetic field anomaly along the pipeline axis, accurately depicts the abrupt change in magnetic circuit at the defect edge, and effectively distinguishes the defect boundary from the smooth background signal fluctuation. The second-order spatial gradient of the normalized anomaly offset value is further calculated to characterize the curvature change of the leakage magnetic field anomaly, accurately locates the abrupt transition zone between the defect center and the defect boundary, enhances the characteristics of defect-induced local magnetic field distortion, and further eliminates slowly changing global operating condition disturbances and background noise interference.

[0047] By fusing multi-dimensional features, the local gradient energy of discrete detection units along each axis is calculated. The normalized anomaly offset amplitude, first-order spatial gradient, second-order spatial gradient and detection step size are weighted and fused to integrate the three types of anomaly features with different dimensions into a single comprehensive anomaly evaluation index, which fully preserves the amplitude, boundary and morphological features of the defect. Then, the preset gradient threshold is used to complete the screening. Detection units that continuously exceed the threshold are merged into anomaly segments, and finally a candidate anomaly field gradient set is generated.

[0048] Based on the candidate anomaly field gradient set, threshold segmentation is performed using local gradient energy as the core indicator to extract continuous anomaly segments and remove single-point noise and short-range spurious peaks. Smoothing fitting and peak search are performed on each effective anomaly segment to accurately determine spatial location parameters such as the start and end coordinates, center position, and equivalent width of the anomaly segment. Combining the local gradient energy of each detection unit within the anomaly segment, a spatial weighted method is used to calculate the comprehensive anomaly intensity of the anomaly segment, highlighting high-energy anomalies at the segment center and suppressing interference from scattered disturbances at the edges. Finally, by using constraints such as reconstruction threshold and minimum continuous length, defect-induced magnetic field disturbance zones are selected.

[0049] All defect-induced magnetic field disturbance regions are mapped to the surface coordinate system of the hydrogen-doped pipeline. The core information such as the spatial coordinates, gradient distribution, and anomaly intensity of each abnormal region are integrated to generate a set of abnormal magnetic field gradient changes and a distribution map of defect-induced magnetic field disturbances. This completes the full-process extraction and spatial representation of deep defect features in the leakage magnetic field channel.

[0050] The method based on the composite detection reference field model performs eddy current response disturbance identification induced by near-surface microcracks in hydrogen-doped pipelines, extracts the abnormal change sequence of near-surface current density, and generates a crack disturbance response feature map, specifically including: Based on the reference electromagnetic induction response matrix and the set of axial discrete detection units, high-frequency electromagnetic excitation parameters and detection coil parameters are configured, and high-frequency alternating electromagnetic excitation is applied to the hydrogen-doped pipe along the detection path to form a skin layer induced current near the surface of the hydrogen-doped pipe. The measured equivalent resistance and measured equivalent inductance of each axial discrete detection unit are collected synchronously. The standard skin depth and measured equivalent impedance amplitude of each axial discrete detection unit are calculated. A measured eddy current complex response sequence is established that corresponds one-to-one with the measured equivalent impedance amplitude and the reference electromagnetic induction response matrix. Based on the standard skin depth and the measured equivalent impedance amplitude, the actual induced current density of each axial discrete detection unit is calculated by inversion, and a comparison sequence of the actual induced current density and the reference current density is generated. The formula for calculating the actual induced current density is as follows: ; In the formula, The actual induced current density of the i-th axial discrete detection unit; Let be the measured equivalent impedance amplitude of the i-th axial discrete detection unit; The measured induced voltage amplitude of the i-th axial discrete detection unit; The equivalent radius of the detection coil; Let be the standard skin depth of the i-th axial discrete detection unit; denoted as the measured phase lag of the i-th axial discrete detection unit.

[0051] Based on the comparison sequence of actual induced current density and reference current density, for each axial discrete detection unit, combined with the detection condition disturbance parameters, the normalized current density offset value and the corresponding first-order spatial gradient and second-order spatial gradient are calculated. By combining the normalized current density offset value and the corresponding first-order and second-order spatial gradients, the local anomaly energy of each axial discrete detection unit is calculated and candidate crack zones are obtained. The comprehensive crack intensity of each candidate crack zone is obtained and the near-surface crack disturbance zone is screened. The near-surface crack disturbance zone is mapped to the surface coordinate system of the hydrogen-doped pipeline to generate the near-surface current density anomaly change sequence and crack disturbance response feature map.

[0052] Based on the reference electromagnetic induction response matrix in the composite detection reference field model and the set of axial discrete detection units, the high-frequency electromagnetic excitation parameters and detection coil parameters are adapted and configured. The excitation frequency matches the detection requirements of near-surface microcracks in hydrogen-doped pipelines and is significantly higher than the quasi-static magnetization frequency of the leakage magnetic field channel, balancing surface detection sensitivity and signal stability. A high-frequency alternating electromagnetic excitation is applied to the pipe wall along a pre-set equidistant detection path along the pipeline axis. Through the electromagnetic induction effect, an induced current layer constrained by the skin effect is formed on the near-surface of the pipeline, focusing the effective range of the electromagnetic field on the near-surface layer where hydrogen-induced damage initiation occurs.

[0053] Simultaneously with high-frequency excitation, the measured equivalent resistance and equivalent inductance of each axial discrete detection unit are acquired through a phase-locked sampling and impedance analysis module. The measured equivalent impedance amplitude of each detection unit is calculated by inversion, and the standard skin depth of the corresponding detection unit is calculated simultaneously to clarify the effective depth boundary of the high-frequency electromagnetic field on the surface of the pipe material. Through the spatial index of the axial discrete detection units, a one-to-one mapping relationship is established between the measured impedance, induced voltage amplitude, phase hysteresis, and other complex response data and the reference electromagnetic induction response matrix. The spatial positions are bound according to the axial sequence of the detection path to form a measured eddy current complex response sequence.

[0054] Based on the standard skin depth, measured equivalent impedance amplitude, induced voltage, and phase information of each detection unit, the actual induced current density near the pipe surface is calculated by inversion. The original voltage and impedance signals are converted into current density quantities that can directly reflect the continuity of the near-surface conductive channel, accurately capturing the cutting and flow disturbance effects of microcracks on the induced current loop. At the same time, the reference current density under the same location and operating conditions is retrieved from the reference electromagnetic induction response matrix to generate a comparison sequence between the actual induced current density and the reference current density.

[0055] Based on the comparison sequence, and combined with operational disturbance parameters such as lift-off fluctuations, temperature fluctuations, and detection speed fluctuations during the detection process, the normalized current density offset value is calculated to eliminate operational interference and extract current density anomalies induced only by microcracks. The first-order and second-order spatial gradients of the offset value are further calculated. The first-order gradient characterizes the current abrupt change at the crack edge, while the second-order gradient enhances the distortion characteristics of the crack center and the transition zones at both ends. By fusing the normalized offset amplitude, first-order spatial gradient, and second-order spatial gradient, the local anomalous energy of each detection unit is calculated. Sections continuously exceeding the threshold are selected through a preset threshold, eliminating single-point noise and short-range spurious peaks to form candidate crack zones.

[0056] The comprehensive crack intensity of each candidate crack region is calculated. A spatial weighting method is used to highlight the high-energy anomaly at the crack region center and suppress the interference of scattered disturbances at the edges. Then, effective near-surface crack disturbance regions are selected through preset crack reconstruction thresholds and minimum continuous length constraints. All near-surface crack disturbance regions are mapped to the surface coordinate system of the hydrogen-doped pipeline. Core information such as the spatial coordinates, boundary range, and anomaly intensity of the anomaly regions is integrated to finally generate a near-surface current density anomaly change sequence and a crack disturbance response feature map, thus accurately extracting the near-surface microcrack features of the eddy current channel.

[0057] The process of fusing and judging multi-source perturbation features by combining the set of abnormal magnetic field gradient changes with the sequence of abnormal near-surface current density changes, and outputting the defect distribution results and defect level assessment results of the hydrogen-doped pipeline, specifically includes: Obtain the set of abnormal magnetic field gradient changes and the sequence of abnormal near-surface current density changes, and establish a unified coordinate system with the detection start point as the axial zero point and the cumulative displacement of the encoder as the axial reference; Map the magnetic flux leakage anomaly region and the eddy current anomaly region to a unified coordinate system, and calculate the axial overlap ratio of each group of magnetic flux leakage anomaly region and eddy current anomaly region. Calculate the center offset distance between each group of magnetic flux leakage anomaly zone and eddy current anomaly zone, and combine the axial overlap ratio to obtain the composite defect alignment set; For each group of composite defect regions in the composite defect alignment set, the perturbation consistency coefficient and composite confidence value are calculated by combining the axial overlap ratio and the center offset distance, and a composite defect confidence sequence is generated. The formula for calculating the disturbance consistency coefficient is as follows: ; In the formula, is the disturbance consistency coefficient between the 0th leakage magnetic anomaly region and the qth eddy current anomaly region; denoted as the axial overlap ratio between the 0th leakage magnetic anomaly region and the qth eddy current anomaly region; The center offset distance between the 0th magnetic flux leakage anomaly region and the qth eddy current anomaly region; The center offset attenuation length; is the normalized flux leakage intensity of the o-th flux leakage anomaly region; is the normalized intensity of the eddy current in the q-th eddy current anomaly region; Understandably, the normalized leakage flux intensity represents the normalized leakage flux intensity, the normalized eddy current intensity represents the normalized eddy current intensity, and so on.

[0058] The formula for calculating the composite credibility value is: ; In the formula, The composite confidence value is the result of the o-th magnetic flux leakage anomaly region and the q-th eddy current anomaly region forming a composite defect. This is the slope control coefficient; This is the confidence threshold.

[0059] Based on the composite defect confidence sequence, for each group of composite defect regions, the corresponding leakage magnetic flux normalized intensity, eddy current normalized intensity, composite confidence value and axial influence length are extracted, and the weighted score value of each group of composite defect regions is calculated. Based on the rating score, the composite defects are classified and the risk level is determined to generate the defect level assessment results for the hydrogen-doped pipeline. The defect level assessment results are mapped to the surface coordinate system of the hydrogen-doped pipeline to generate the defect distribution results of the hydrogen-doped pipeline.

[0060] The abnormal magnetic field gradient change set output by the leakage magnetic field channel and the abnormal near-surface current density change sequence output by the eddy current channel are obtained. A unified spatial coordinate system is established with the detection start point as the axial zero point and the cumulative displacement of the detection head encoder as the axial reference. At the same time, the circumferential development line of the pipe is used as the circumferential reference, and the array channel number or circumferential angle calibration value is used as the circumferential coordinate reference to eliminate the spatial index deviation between the two detection channels. If there is a difference in the sampling step size used for leakage magnetic field and eddy current detection, linear interpolation or cubic spline interpolation is used to resample the two types of abnormal data to the same axial interval to ensure that the spatial coordinates of the leakage magnetic field anomaly area and the eddy current anomaly area are completely aligned.

[0061] The defect-induced magnetic field disturbance region of the magnetic flux leakage channel and the near-surface crack disturbance region of the eddy current channel are all mapped to a unified spatial coordinate system. The axial overlap ratio of the magnetic flux leakage anomaly region and the eddy current anomaly region are calculated group by group to quantify the degree of spatial co-location of the two types of anomalies. The higher the overlap ratio, the greater the probability that the two types of anomalies originate from the same physical defect region. The center offset distance of each group of magnetic flux leakage anomaly regions and eddy current anomaly regions is calculated simultaneously to identify pseudo-matches with slight boundary overlap but obvious misalignment of the main peak of the anomaly, thus avoiding the erroneous fusion of non-homogeneous anomalies. By using preset axial overlap ratio thresholds and center offset distance thresholds, anomaly region pairs that simultaneously meet the spatial co-location requirements are selected to form a composite defect alignment set.

[0062] For each pair of anomalous regions in the composite defect alignment set, the leakage magnetic flux anomaly intensity and eddy current anomaly intensity are first normalized globally, mapping the anomalous intensities of both channels to a dimensionless range of 0 to 1, thus eliminating dimensional differences between different detection channels. Combining the axial overlap ratio, center offset distance, and dual-channel normalized intensity, a perturbation consistency coefficient is calculated. This coefficient simultaneously considers spatial co-location, center offset penalty, and dual-channel energy balance, effectively suppressing pseudo-composite situations where one channel exhibits significant anomalies while the other has no response. A high consistency coefficient can only be obtained when the two types of anomalies highly overlap spatially and both channel anomaly intensities are relatively significant. The consistency coefficient is then converted into a composite confidence value in the 0-1 range. Through slope control coefficients and confidence thresholds, the probability of composite defects is standardized and quantified. Finally, a composite defect confidence sequence is generated according to the anomalous region pairs.

[0063] Based on the composite defect confidence sequence, for each group of valid composite defect areas, four core evaluation dimensions are extracted: normalized flux leakage intensity, normalized eddy current intensity, composite confidence value, and axial influence length. These correspond to the degree of deep section loss, the degree of near-surface crack activity, the consistency of dual-channel identification, and the scale of defect spatial expansion, respectively. Through weighted fusion calculation with preset weights, a grade score for each group of composite defect areas is obtained. Based on the numerical range of the grade score, composite defects are divided into three risk levels: minor anomaly, medium-level composite defect, and high-risk composite defect. For anomaly areas with a composite confidence value below a preset lower limit, even if the score value is high, they are not included in the composite defect category and are separately marked as single-channel anomaly areas awaiting review, ultimately forming a complete defect grade assessment result for hydrogen-doped pipelines.

[0064] The defect level assessment results are remapped to the surface coordinate system of the hydrogen-blended pipeline, clarifying the axial and circumferential coordinates, boundary range, defect type, impact length, risk level, and review priority of each defect area. The defects are sorted from high to low risk level to generate the defect distribution results and corresponding defect list of the hydrogen-blended pipeline, forming the final inspection conclusion that can be directly used for pipeline operation and maintenance, integrity assessment, and maintenance decisions.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting hydrogen-doped pipelines based on a combined eddy current and magnetic flux leakage detection system, characterized in that, The method includes: Obtain the reference magnetic field distribution vector and reference electromagnetic induction response matrix of the hydrogen-doped pipeline under complete state, and construct a composite detection reference field model of the hydrogen-doped pipeline. Based on the composite detection reference field model, the leakage magnetic field response acquisition and abnormal field reconstruction under the defect disturbance of hydrogen-doped pipeline are performed, the abnormal magnetic field gradient change set is extracted, and the defect-induced magnetic field disturbance distribution map is generated. Based on the composite detection reference field model, the eddy current response disturbance identification induced by near-surface microcracks in hydrogen-doped pipelines is performed, the abnormal change sequence of near-surface current density is extracted, and the crack disturbance response feature map is generated. The set of abnormal magnetic field gradient changes and the sequence of abnormal near-surface current density changes are fused and judged by multi-source perturbation features, and the defect distribution results and defect level assessment results of hydrogen-doped pipelines are output.

2. The method for detecting hydrogen-doped pipelines based on combined eddy current and magnetic flux leakage detection according to claim 1, characterized in that, The process of obtaining the reference magnetic field distribution vector and reference electromagnetic induction response matrix of the hydrogen-doped pipeline under its complete state, and constructing a composite detection reference field model for the hydrogen-doped pipeline, includes the following steps: Obtain the base material information, geometric parameters and electromagnetic parameters of the material in the complete state of the hydrogen-doped pipeline. Establish equidistant detection paths along the axial direction of the hydrogen-doped pipeline and divide them into a set of axial discrete detection units. Calculate the equivalent permeability of each axial discrete detection unit. Based on the set of axial discrete detection units and the equivalent permeability of each axial discrete detection unit, combined with the magnetization device parameters and probe lift-off height, the reference magnetic induction intensity of each axial discrete detection unit under the complete state of the hydrogen-doped pipeline is calculated, and the reference magnetic field distribution vector is generated. Based on the set of axial discrete detection units and the equivalent permeability of each axial discrete detection unit, combined with the high-frequency excitation parameters and detection coil parameters, the standard skin depth and reference induced voltage amplitude of each axial discrete detection unit under the complete state of the hydrogen-doped pipeline are calculated, and the reference electromagnetic induction response matrix is ​​generated. Based on the reference magnetic field distribution vector and the reference electromagnetic induction response matrix, an excitation scan of the hydrogen-doped pipeline in its complete state is performed, the composite reference response value of each axial discrete detection unit is calculated, and a composite detection reference field model containing the reference magnetic field distribution vector and the reference electromagnetic induction response matrix is ​​generated.

3. The method for detecting hydrogen-doped pipelines based on combined eddy current and magnetic flux leakage detection according to claim 1, characterized in that, The process, based on a composite detection reference field model, involves acquiring the leakage magnetic field response and reconstructing the abnormal field under hydrogen-doped pipeline defect disturbances, extracting the abnormal magnetic field gradient change set, and generating a defect-induced magnetic field disturbance distribution map. Specifically, this includes: Based on the reference magnetic field distribution vector and the set of axial discrete detection units, the working parameters of the magnetization device and the magnetic sensing unit are configured to apply a constant magnetization field to the hydrogen-doped pipeline until the wall of the hydrogen-doped pipeline enters a stable near-saturation magnetization range. Along the detection path of the hydrogen-doped pipeline, the magnetic sensing unit collects the three components of leakage magnetic field signals in the axial, radial and tangential directions of each axial discrete detection unit. Combined with the detection propulsion speed and sampling time, the axial sampling position corresponding to each sampling point is calculated. Based on the three components of leakage magnetic field signals in the axial, radial and tangential directions of each axial discrete detection unit, the measured total leakage magnetic field amplitude of each axial discrete detection unit is synthesized. The measured total leakage magnetic flux amplitude of each axial discrete detection unit is bound to the corresponding axial sampling position to form a measured leakage magnetic flux response sequence arranged continuously along the detection path.

4. The method for detecting hydrogen-doped pipelines based on combined eddy current and magnetic flux leakage detection according to claim 3, characterized in that, Based on the measured leakage magnetic response sequence, the reference magnetic induction intensity corresponding to each axial discrete detection unit is retrieved, and the measured-reference correspondence under the same position and working condition constraints is established. Combined with the detection working condition disturbance parameters, the normalized abnormal offset value of each axial discrete detection unit is calculated. The formula for calculating the normalized outlier offset is as follows: ； In the formula, is the normalized anomaly offset value of the i-th axial discrete detection unit; The measured total leakage flux amplitude of the i-th axial discrete detection unit; The reference magnetic flux density is the i-th axial discrete detection unit. For the local calibration correction coefficient of the i-th axial discrete detection unit; The actual lift-off height between the magnetization module and the outer surface of the hydrogen-doped pipe at the i-th axial discrete detection unit; Let be the detection temperature of the i-th axial discrete detection unit; To calibrate the detection temperature; Let be the magnetization current of the i-th axial discrete detection unit; To calibrate the magnetizing current; This is the compensation coefficient; A local sliding window is set based on the normalized abnormal offset value, and the local offset ratio of each axial discrete detection unit is calculated. The formula for calculating the local offset ratio is as follows: ； In the formula, The local offset ratio of the i-th axial discrete detection unit; To prevent stable terms with a denominator of zero; This represents the number of axial discrete detection units within the local sliding window. It is the sum of the absolute values ​​of the normalized anomaly offset values ​​of each axial discrete detection unit within a local sliding window centered on the i-th axial discrete detection unit; The normalized abnormal offset value of each axial discrete detection unit is bound to the local offset ratio to generate an initial abnormal magnetic field offset sequence.

5. The method for detecting hydrogen-doped pipelines based on combined eddy current and magnetic flux leakage detection according to claim 4, characterized in that, Based on the initial abnormal magnetic field offset sequence, the normalized abnormal offset values ​​of each axial discrete detection unit are sorted according to the axial position of the detection path. The first-order spatial gradient of the normalized anomaly offset value of each axial discrete detection unit is calculated using the central difference method. Based on the first-order spatial gradient, the second-order spatial gradient of the normalized anomaly offset value of each axial discrete detection unit is calculated. By combining the normalized anomaly offset value and the corresponding first-order spatial gradient, second-order spatial gradient and detection step size, the local gradient energy of each axial discrete detection unit is calculated to generate a candidate anomaly field gradient set.

6. The method for detecting hydrogen-doped pipelines based on combined eddy current and magnetic flux leakage detection according to claim 5, characterized in that, Obtain the set of candidate anomaly field gradients, and filter out anomaly segments based on local gradient energy; Smoothing fitting and peak search are performed on each abnormal section to determine the location parameters of the abnormal section. Combined with the local gradient energy of each axial discrete detection unit in the abnormal section, the comprehensive abnormal intensity of the abnormal section is calculated and the defect-induced magnetic field disturbance zone is screened out. The formula for calculating the overall anomaly intensity is as follows: ； In the formula, The overall anomaly intensity of the k-th anomalous segment; The number of axial discrete detection units contained in the k-th abnormal segment; Let be the local gradient energy of the i-th axial discrete detection unit; The axial position of the i-th axial discrete detection unit within the abnormal section; This is the center location of the k-th abnormal segment; The equivalent width of the k-th abnormal segment; All defect-induced magnetic field disturbance regions are mapped to the surface coordinate system of the hydrogen-doped pipeline to generate a set of abnormal magnetic field gradient changes and a distribution map of defect-induced magnetic field disturbances.

7. The method for detecting hydrogen-doped pipelines based on combined eddy current and magnetic flux leakage detection according to claim 1, characterized in that, The method based on the composite detection reference field model performs eddy current response disturbance identification induced by near-surface microcracks in hydrogen-doped pipelines, extracts the abnormal change sequence of near-surface current density, and generates a crack disturbance response feature map, specifically including: Based on the reference electromagnetic induction response matrix and the set of axial discrete detection units, high-frequency electromagnetic excitation parameters and detection coil parameters are configured, and high-frequency alternating electromagnetic excitation is applied to the hydrogen-doped pipe along the detection path to form a skin layer induced current near the surface of the hydrogen-doped pipe. The measured equivalent resistance and measured equivalent inductance of each axial discrete detection unit are collected synchronously. The standard skin depth and measured equivalent impedance amplitude of each axial discrete detection unit are calculated. A measured eddy current complex response sequence is established that corresponds one-to-one with the measured equivalent impedance amplitude and the reference electromagnetic induction response matrix. Based on the standard skin depth and the measured equivalent impedance amplitude, the actual induced current density of each axial discrete detection unit is calculated by inversion, and a comparison sequence of the actual induced current density and the reference current density is generated. The formula for calculating the actual induced current density is as follows: ； In the formula, The actual induced current density of the i-th axial discrete detection unit; Let be the measured equivalent impedance amplitude of the i-th axial discrete detection unit; The measured induced voltage amplitude of the i-th axial discrete detection unit; The equivalent radius of the detection coil; Let be the standard skin depth of the i-th axial discrete detection unit; The measured phase lag of the i-th axial discrete detection unit; The actual lift-off height between the magnetization module and the outer surface of the hydrogen-doped pipe at the i-th axial discrete detection unit.

8. The method for detecting hydrogen-doped pipelines based on combined eddy current and magnetic flux leakage detection according to claim 7, characterized in that, Based on the comparison sequence of actual induced current density and reference current density, for each axial discrete detection unit, combined with the detection condition disturbance parameters, the normalized current density offset value and the corresponding first-order spatial gradient and second-order spatial gradient are calculated. By combining the normalized current density offset value and the corresponding first-order and second-order spatial gradients, the local anomaly energy of each axial discrete detection unit is calculated and candidate crack zones are obtained. The comprehensive crack intensity of each candidate crack zone is obtained and the near-surface crack disturbance zone is screened. The near-surface crack disturbance zone is mapped to the surface coordinate system of the hydrogen-doped pipeline to generate the near-surface current density anomaly change sequence and crack disturbance response feature map.

9. The method for detecting hydrogen-doped pipelines based on combined eddy current and magnetic flux leakage detection according to claim 1, characterized in that, The process of fusing and judging multi-source perturbation features by combining the set of abnormal magnetic field gradient changes with the sequence of abnormal near-surface current density changes, and outputting the defect distribution results and defect level assessment results of the hydrogen-doped pipeline, specifically includes: Obtain the set of abnormal magnetic field gradient changes and the sequence of abnormal near-surface current density changes, and establish a unified coordinate system with the detection start point as the axial zero point and the cumulative displacement of the encoder as the axial reference; Map the magnetic flux leakage anomaly region and the eddy current anomaly region to a unified coordinate system, and calculate the axial overlap ratio of each group of magnetic flux leakage anomaly region and eddy current anomaly region. Calculate the center offset distance between each group of magnetic flux leakage anomaly zone and eddy current anomaly zone, and combine the axial overlap ratio to obtain the composite defect alignment set; For each group of composite defect regions in the composite defect alignment set, the perturbation consistency coefficient and composite confidence value are calculated by combining the axial overlap ratio and the center offset distance, and a composite defect confidence sequence is generated. The formula for calculating the disturbance consistency coefficient is as follows: ； In the formula, is the disturbance consistency coefficient between the 0th leakage magnetic anomaly region and the qth eddy current anomaly region; denoted as the axial overlap ratio between the 0th leakage magnetic anomaly region and the qth eddy current anomaly region; The center offset distance between the 0th magnetic flux leakage anomaly region and the qth eddy current anomaly region; The center offset attenuation length; is the normalized flux leakage intensity of the o-th flux leakage anomaly region; is the normalized intensity of the eddy current in the q-th eddy current anomaly region; To prevent stable terms with a denominator of zero; The formula for calculating the composite credibility value is: ； In the formula, The composite confidence value is the result of the o-th magnetic flux leakage anomaly region and the q-th eddy current anomaly region forming a composite defect. This is the slope control coefficient; This is the confidence threshold.

10. The method for detecting hydrogen-doped pipelines based on combined eddy current and magnetic flux leakage detection according to claim 9, characterized in that, Based on the composite defect confidence sequence, for each group of composite defect regions, the corresponding leakage magnetic flux normalized intensity, eddy current normalized intensity, composite confidence value and axial influence length are extracted, and the weighted score value of each group of composite defect regions is calculated. Based on the rating score, the composite defects are classified and the risk level is determined to generate the defect level assessment results for the hydrogen-doped pipeline. The defect level assessment results are mapped to the surface coordinate system of the hydrogen-doped pipeline to generate the defect distribution results of the hydrogen-doped pipeline.

Images